The Agent Skills Directory

[COMMAND_EXECUTION]: The skill workflow involves executing git commands such as git status, git diff, and git commit to manage repository changes as defined in SKILL.md and guide.md.
[COMMAND_EXECUTION]: The 'Reality Check' section in guide.md instructs the agent to run uds check (a tool from the vendor asiaostrich) and npx commitlint to validate commit message standards.
[EXTERNAL_DOWNLOADS]: The skill uses npx in guide.md to fetch the commitlint package from the npm registry, which is a well-known service registry.
[REMOTE_CODE_EXECUTION]: The validation step utilizes npx to dynamically execute the commitlint tool from a remote source.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from the files it processes.
Ingestion points: The skill reads untrusted data from the local repository via git diff --staged and git status as specified in SKILL.md.
Boundary markers: None. No delimiters or specific instructions are provided to the agent to ignore or isolate content within the git diffs.
Capability inventory: The skill can modify the repository history using git commit and execute shell commands through git and npx.
Sanitization: There is no evidence of sanitization or escaping applied to the content of the code diffs before they are processed by the LLM to generate the commit message.

commit